Published ahead of print on April 7, 2005, doi:10.1164/rccm.200411-1587OC
© 2005 American Thoracic Society doi: 10.1164/rccm.200411-1587OC
Antiproliferative Effects of Phosphodiesterase Type 5 Inhibition in Human Pulmonary Artery CellsSection on Experimental Medicine and Toxicology, Imperial College London, Hammersmith Hospital, London; and Pfizer Global Research and Development, Kent, United Kingdom Correspondence and requests for reprints should be addressed to John Wharton, Ph.D., Section on Experimental Medicine and Toxicology, Imperial College London, Hammersmith Campus, Du Cane Road, London W12 ONN, UK. E-mail: j.wharton{at}imperial.ac.uk
Rationale: Phosphodiesterase Type 5 (PDE5) inhibition represents a novel strategy for the treatment of pulmonary hypertension. Objectives: Our aim was to establish the distribution of PDE5 in the pulmonary vasculature and effects of PDE5 inhibition on pulmonary artery smooth muscle cells (PASMCs). Methods and Measurements: PDE5 expression was examined by immunohistochemistry and Western blotting, in both normal and hypertensive lung tissues. DNA synthesis, proliferation, PDE activity, and apoptosis were measured in distal human PASMCs treated with soluble guanylyl cyclase activators (nitric oxide donors and BAY41–2272) and sildenafil. Main Results: Cells containing PDE5 and  -smooth muscle actin occurred throughout the pulmonary vasculature, including obstructive intimal lesions. Three molecular forms of PDE5 were identified and protein expression was greater in hypertensive than control lung tissue. Most cyclic guanosine monophosphate hydrolysis (about 80%) in cultured cells was attributed to PDE5. Sildenafil induced a greater elevation of intracellular cyclic guanosine monophosphate levels compared with nitric oxide donors and BAY41–2272 (about 10-fold versus about 2-fold) and cotreatment had a synergistic effect, increasing cyclic nucleotide levels up to 50-fold. Dual stimulation of soluble guanylyl cyclase and inhibition of PDE5 activities also had significant downstream effects, increasing phosphorylation of vasodilator-stimulated phosphoprotein, reducing DNA synthesis and cell proliferation, and stimulating apoptosis, and these effects were mimicked by cyclic guanosine monophosphate analogs. Conclusions: Phosphodiesterase Type 5 is the main factor regulating cyclic guanosine monophosphate hydrolysis and downstream signaling in human PASMCs. The antiproliferative effects of this signaling pathway may be significant in the chronic treatment of pulmonary hypertension with PDE5 inhibitors such as sildenafil.
Key Words: cell proliferation • cyclic nucleotides • hypertension, pulmonary Pulmonary arterial hypertension (PAH) is a progressive disease of pulmonary arteries that is characterized by a sustained increase in pulmonary pressure and vascular remodeling (1). The structural changes typically involve muscularization and thickening of precapillary pulmonary arteries, intimal proliferation, obliterative lesions, and thrombosis in situ. Many of these changes can be attributed to the proliferation of pulmonary artery smooth muscle cells (PASMCs) or myofibroblasts and attention has therefore focused on the potential antiproliferative role of therapeutic agents (1). Nitric oxide (NO) reduces pulmonary vascular resistance by stimulating soluble guanylyl cyclase (sGC), thereby increasing cyclic guanosine monophosphate (cGMP) levels and stimulating cGMP-dependent protein kinase (PKG) in pulmonary vascular smooth muscle, and several therapies have emerged that target this pathway (1, 2). In addition to regulating vascular tone, NO and cGMP signaling can also inhibit proliferation and induce apoptosis in vascular smooth muscle cells (SMCs) (3–6). The mechanisms involved are still unclear; however, some reports have suggested that the antiproliferative responses to NO and cGMP in systemic vascular SMCs are dependent on cross-talk with the cAMP signaling pathway and activation of cAMP-dependent protein kinase (7). Studies to date have also examined the effects of NO and cGMP on the growth of PASMCs from experimental animals (3, 5), rather than cells isolated from human pulmonary arteries, which may differ in their cGMP content and signaling (8). Intracellular cGMP levels are determined by the rate of cGMP hydrolysis as well as synthesis, and cGMP-specific phosphodiesterase (PDE) Type 5 (PDE5) and calcium/calmodulin-dependent PDE Type 1 (PDE1) are the major enzymes responsible for degrading cGMP in vascular SMCs (9). PDE5 is mainly responsible for cGMP hydrolysis in the lung (10) and lobar pulmonary arteries (11–13) and studies indicate that the PDE5 inhibitor sildenafil is effective in reducing pulmonary vascular resistance in hypoxia-induced pulmonary hypertension (14–16) and in patients with severe pulmonary hypertension (17–20). Furthermore, studies in animal models of pulmonary hypertension suggest that chronic PDE5 inhibition attenuates vascular remodeling as well as the rise in pulmonary artery pressure (21–24). Chronic inhibition of PDE5 activity, by drugs such as sildenafil, therefore represents an attractive therapeutic strategy for manipulating both pulmonary vascular structure and tone. We hypothesized that PDE5 protein expression may be increased in the remodeled and hypertensive pulmonary vasculature and that PDE5 activity is a critical factor regulating human PASMC proliferation and apoptosis. To date, however, no study has established the distribution of PDE5 in remodeled pulmonary vessels of the hypertensive human lung and it is not known whether PDE5 inhibitors and activators of cGMP signaling have the ability to influence human PASMC proliferation and vascular remodeling. We therefore sought to (1) determine the distribution of PDE5 immunoreactivity in normal and hypertensive lung tissues, (2) determine the effects of PDE5 inhibition and cGMP stimulation on proliferation and apoptosis in PASMCs derived from small human pulmonary arteries, and (3) investigate whether responses to PDE5 inhibition and cGMP stimulation involve cross-talk between cGMP and cAMP signaling. Some of the results of these studies have been previously reported in the form of an abstract (25).
See the online supplement for further details concerning methods and reagents.
Lung Tissues
Immunohistochemistry and Western Blotting
Phosphodiesterase Activity and Intracellular Cyclic Nucleotide Levels
DNA Synthesis, Cell Proliferation, and Apoptosis
Statistical Analysis
Distribution of PDE5 Immunoreactivity in Pulmonary Vessels PDE5 immunoreactivity was localized to smooth muscle cells in elastic and muscular pulmonary arteries, pulmonary veins, and bronchial blood vessels, as well as the muscle layer in both bronchi and bronchioles (Figures 1A–1E). Both PDE5 antisera provided similar patterns of immunostaining and PDE5A1 immunostaining was abolished after preabsorption of the primary antisera with peptide antigen (0.1–1 µg/ml) (Figure 1F). In lung tissues from patients with either idiopathic PAH or PAH associated with congenital heart disease, PDE5-immunoreactive smooth muscle cells were distributed throughout the remodeled pulmonary vasculature, including intimal plaques in elastic arteries, the hypertrophied medial layer, and obstructive intimal lesions in small arteries and the neomuscularized distal microvasculature (Figures 2A–2D).
Western blotting demonstrated three distinct bands of PDE5 immunoreactivity in lung homogenates with molecular sizes of 98 kD (PDE5A1), 86 kD, and 77 kD (Figure 3A). PDE5 immunoreactivity, normalized to total protein content, was increased in the pulmonary hypertensive lung, the expression of the predominant 98- and 77-kD molecular forms being significantly greater (p < 0.05) in lung extracts from pulmonary hypertensive patients compared with samples from control subjects (Figures 3A and 3B). Vascular remodeling in pulmonary hypertensive lung was also associated with an increase in -SMA immunoreactivity in tissue extracts (Figures 3A and 3B).
Intracellular Cyclic Nucleotide Levels in PASMCs Under basal conditions, low levels of cGMP were found in PASMCs. Sildenafil raised intracellular cGMP levels in a time- and concentration-dependent manner, inducing increases of 5- to 10-fold (Figures 4A and 4B; and see Figures E1 and E2 in the online supplement). Stimulation of sGC activity, by either NO donors (DETA NONOate and S-nitro-N-acetylpenicillamine [SNAP]) or the NO-independent sGC activator BAY41–2272, induced only a modest rise (up to twofold) in intracellular cGMP, which was attenuated by the sGC inhibitor oxadiazole[4,3-a]quinoxalin-1-one (ODQ). In contrast, dual sGC stimulation and PDE5 inhibition had a synergistic effect, increasing cGMP levels up to 50-fold (Figures 4C and 4D; and Figures E1 and E2). The effects of NO donors and PDE5 inhibitors on cGMP levels were not accompanied by significant changes in intracellular cAMP levels and the PDE1 inhibitor 8-methoxy-methyl-3-isobutyl-1-methylxanthine (8MM-IBMX) had a significant effect on cGMP levels only at the highest concentration used (Figures 4E and 4F).
PDE Activity in PASMCs cGMP- and cAMP-PDE activities were determined in both cytosolic and membrane cell fractions (Figures 5A and 5B). The total cGMP-PDE activity in these subcellular fractions (514.4 ± 34.9 pmol/minute per milligram protein) was significantly greater than cAMP-PDE activity (88.2 ± 9.7 pmol/minute per milligram protein; n = 5 isolates; p < 0.0001) and was equally distributed between the cytosol (52 ± 2%) and membrane fraction (48 ± 2%). Sildenafil (10–6 M) inhibited most of the cGMP-PDE activity, in both cytosolic (83 ± 3%; n = 10) and membrane fractions (87 ± 2%; n = 8), and the general PDE inhibitor 3-isobutyl-1-methylxanthine (IBMX) also reduced cGMP hydrolysis to a similar extent, indicating that PDE5 was responsible for most of the cGMP degradation. This was supported by the finding that only a minor proportion of the cGMP- and cAMP-PDE activities (13–16%) was inhibited by EGTA and therefore attributable to PDE1. However, substrate-dependent differences were observed regarding the effect of Ca2+/CM stimulation and response to selective PDE1 inhibitors (Figures 5A–5C). Thus, whereas cGMP hydrolytic activity was relatively insensitive to Ca2+/CM stimulation of cAMP hydrolysis was markedly enhanced, PDE1 activity increasing in the cytosolic and membrane fractions by 4.7- and 20.2-fold, respectively (Figures 5A and 5B). The PDE1 inhibitor 8MM-IBMX was also more potent at inhibiting cAMP hydrolysis than cGMP-PDE activity (Figure 5C). cGMP hydrolysis was attenuated by PDE5 and PDE1 inhibitors, as well as by IBMX, but not by PDE2-, PDE3-, or PDE4-selective inhibitors.
Both subcellular fractions displayed cAMP-PDE activity, the cytosol containing more activity than the membrane fraction (56.4 ± 6.4 versus 31.8 ± 3.9 pmol/minute per milligram protein; n = 5; p < 0.001). IBMX inhibited cAMP hydrolysis in both the cytosolic (103 ± 3%) and membrane fractions (82 ± 5%) and a proportion was attributed to PDE1, PDE2, PDE3, and PDE4 after inhibition experiments with EGTA, erythro-9-(2-hydroxy-3-nonyl)adenine (EHNA), cilostamide, and rolipram, respectively. Expression of PDE3 activity in PASMC extracts was confirmed by the characteristic inhibition of cAMP hydrolysis in the presence of micromolar concentrations of exogenous cGMP (Figure 5D). However, cAMP-PDE activity was not attenuated after the synergistic elevation of intracellular cGMP levels in cells cotreated with sildenafil and SNAP or BAY41–2272 and, except when sildenafil was used at high concentration (10–5 M), the treatment of cells with PDE5 inhibitors did not generally reduce cAMP-PDE activity (Figure 5E). The sGC activator BAY41–2272 also had no apparent effect on PDE activity at concentrations up to 10–5 M. Inhibition studies on enzyme activity in cell extracts indicated that sildenafil had about 50-fold greater selectivity for cGMP-PDE activity (IC50, 9.14 x 10–8 M; 95% CI, [3.89–21.3] x 10–8 M; n = 6 isolates) over cAMP-PDE activity (IC50, 4.67 x 10–6 M; 95% CI, [0.83–26.3] x 10–6 M; n = 4; p < 0.001). The structurally distinct PDE5 inhibitor T1032 also displayed a relatively low inhibitory potency (IC50, 9.46 x 10–6 M; 95% CI, [2.26–39.5] x 10–6 M; n = 4) for cAMP-PDE activity (Figure 5F).
Western Blotting and VASP Phosphorylation in PASMCs
Phosphorylation of VASP was not observed without prior stimulation of PKG activity. The cGMP analog 8-bromo-cGMP (8Br-cGMP) induced phosphorylation at Ser-239 in a concentration- and time-dependent manner and this was detectable for at least 24 hours after cells were treated (Figure 6B). Stimulation with 8Br-cGMP also resulted in a molecular size shift of VASP, from 46 to 50 kD (Figure 6B). In addition to cGMP analogs, phosphorylation of VASP at Ser-239 was induced by NO donors and BAY41–2272 and the response was potentiated by sildenafil (Figure 6C).
DNA Synthesis, Cell Proliferation, and Apoptosis in PASMCs
Serum withdrawal induced a characteristic apoptotic response, as demonstrated by nuclear condensation with Hoechst 33342 staining and by DNA fragmentation (Figures 8A and 8B). The proportion of apoptotic cells increased after costimulation with SNAP and sildenafil, but did not change with either agent alone (Figure 8A). Treatment with 8Br-cGMP (10–8 to 10–6 M) also induced a concentration-dependent increase in the proportion of cells displaying nuclear condensation. Serum withdrawal and treatment of cells with SNAP and sildenafil or 8Br-cGMP also induced DNA fragmentation, supporting the contention that cell death was mediated via cGMP (Figure 8B). The effects of sildenafil and sGC activators on PASMC proliferation and apoptosis were reproducible between different cell isolates, irrespective of whether they were derived from normal or diseased lung tissues.
Our results provide evidence supporting an independent role for PDE5 and cGMP signaling in regulating the proliferation of distal human PASMCs. Several studies have demonstrated that PDE5 activity contributes to the pathophysiology of pulmonary hypertension by limiting cGMP-mediated vasodilator effects in the pulmonary vasculature (18). Increased PDE5 expression and enzyme activity have been reported in several experimental models of pulmonary hypertension (24, 31–33) and may contribute to the impairment of vasodilator responses in the hypoxic lung (34). Studies in experimental animals have also demonstrated that oral treatment with sildenafil significantly reduces neomuscularization in both hypoxia and monocrotaline models of pulmonary hypertension (21–23) and limits the increase in pulmonary artery medial thickness induced by a left-to-right arterial shunt (24). Sildenafil has been shown to selectively reduce pulmonary vascular resistance in patients with different forms of pulmonary hypertension (17–20) and to reduce right ventricular mass in patients with PAH (35, 36); however, its effect on vascular remodeling in the hypertensive human lung is uncertain.
PDE5 mRNA is widely expressed in human tissues and cells, including the lung and pulmonary SMCs (37). Three splice variants have been identified (PDE5A1, PDE5A2, and PDE5A3), which encode proteins with similar cGMP catalytic activities and sensitivity to sildenafil, but distinct N-terminal sequences, and whereas PDE5A1 and PDE5A2 variants occur in most tissues expression of the smallest splice variant, PDE5A3, seems to be limited to smooth muscle cells (29). We identified three molecular forms of PDE5 immunoreactivity in lung homogenates and isolated PASMCs, these being smaller in size than the molecular mass predicted from cDNA sequences of PDE5 splice variants (9, 29). Expression of PDE5 and Intracellular cGMP levels were increased both by stimulating synthesis and reducing hydrolysis. PDE5 inhibitors were more effective than either NO donors or BAY41–2272 in raising the cGMP content of PASMCs and the levels continued to rise in sildenafil-treated cells, suggesting an underlying basal stimulation of sGC activity. However, even in the absence of sildenafil, NO donors and BAY41–2272 induced downstream signaling, as demonstrated by VASP phosphorylation at Ser-239 and antiproliferative responses. VASP phosphorylation at Ser-239 is an established biochemical marker of PKG activation (38, 39) and has been shown to be involved in the antiproliferative effects of NO and cGMP signaling in aortic SMCs (40). In the present study, VASP Ser-239 phosphorylation was induced both by stimulating cGMP synthesis, with NO donors or BAY41–2272, and via direct PKG activation with stable cGMP analogs. Sildenafil induced VASP Ser-239 phosphorylation, as well as antiproliferative responses, but with less apparent efficiency than direct sGC or PKG activation. In serum-stimulated cell growth experiments, this is likely to reflect, at least in part, a reduction in free drug levels due to serum protein binding (41). Sildenafil was more effective when combined with sGC activators, potentiating the effects of NO donors on VASP phosphorylation, DNA synthesis, and cell proliferation and inducing an apoptotic response, which was again mimicked by 8Br-cGMP. In vivo, pulmonary cGMP production is dependent on guanylyl cyclase stimulation by natriuretic peptides as well as NO and both pathways contribute to the effect of sildenafil in hypoxia-induced pulmonary hypertension and cardiovascular remodeling (14, 42). Continuous stimulation of cGMP production may, however, lead to the accumulated phosphorylation and activation of PDE5 by PKG in human smooth muscle cells (39) and possibly to an increase in PDE5 expression (43, 44). It has also been postulated that PDE5 inhibitors reduce the proliferation of SMCs, such as those isolated from bovine coronary arteries, via competition between cGMP and cAMP for PDE3-dependent hydrolysis and the subsequent stimulation of cAMP signaling (7). Indeed, we showed that exogenous cGMP attenuated cAMP hydrolytic activity in the subcellular fractions of distal human PASMCs, but it is also recognized that intracellular cyclic nucleotide levels vary between species and that endogenous cGMP levels in human SMCs are relatively low (8). Thus, even after treatment with cGMP-elevating agents, the cGMP content of PASMCs was several orders of magnitude lower than that of cAMP and therefore unlikely to compete significantly with cAMP as a substrate for PDE3-dependent hydrolysis. This contention is supported by the lack of an inhibitory effect on cAMP-PDE activity and absence of a corresponding change in cAMP levels in cells treated with cGMP-elevating agents. Our data indicate that PDE5 is the main factor responsible for the hydrolysis of cGMP in distal human PASMCs. This is consistent with previous reports that attributed cGMP hydrolysis in lung tissue (10) and lobar pulmonary arteries to cGMP-specific PDE (PDE5) activity (11–13), but contrasts with studies on cGMP hydrolysis in systemic vessels and SMCs, where a significant proportion of cGMP hydrolysis is also mediated via Ca2+/CM-dependent PDE (PDE1) activity (10, 45). Indeed, distal human PASMCs exhibited relatively little PDE1-dependent cGMP hydrolysis, but PDE1-dependent hydrolysis of cAMP was markedly induced by Ca2+/CM stimulation and represented the predominant cAMP-hydrolytic activity demonstrated in both subcellular fractions. The additional inhibitory effect of sildenafil on cAMP-PDE activity is therefore potentially important, sildenafil having been reported to possess an 80-fold greater selectivity for PDE5 over PDE1 in human tissues (46). In the present study, sildenafil displayed an approximately 50-fold selectivity for cGMP-PDE (PDE5) activity over cAMP-PDE activity in PASMCs, but the maximum plasma concentration ([1.9–8.4] x 10–7 M) achieved at therapeutic doses (25–100 mg) of sildenafil (47) may be too low to significantly inhibit pulmonary cAMP-PDE activity in vivo. In conclusion, this study has demonstrated that PDE5 is expressed in distal PASMCs and remodeled pulmonary arteries, providing a target for PDE5 inhibitors in pulmonary hypertension. Stimulation of the cGMP pathway inhibited DNA synthesis and cell proliferation and promoted apoptosis in isolated distal human PASMCs. These effects were potentiated by sildenafil and appear to be independent of cAMP signaling. This in vitro study does not directly address the question of whether PDE5 inhibition is capable of modifying vascular remodeling in vivo; nonetheless, chronic PDE5 inhibition may have an antiproliferative effect in patients with severe pulmonary hypertension.
Supported by the British Heart Foundation (project grants PG2001041 and PG2001102) and by a Biotechnology and Biological Sciences Research Council CASE Studentship (E.J.G.) with Pfizer Global Research and Development. This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Conflict of Interest Statement: J.W. received £1,750 from Pfizer for lecture fees and attending conferences, and his institution received a grant of £7,765 per annum for 2002–2004 from Pfizer to help finance a BBSRC-CASE Ph.D. studentship for E.J.G. J.W.S. received £1,000 from Pfizer for attending conferences. G.M.O.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. E.J.G. received a bursary of £4,565 per annum during 2002–2004 from Pfizer towards a stipend for a BBSRC-CASE Ph.D. studentship. X.R. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. A.P.F. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. S.C.P. is a full-time employee of Pfizer (UK), holds Pfizer shares and stock options, and his spouse also has shares in Pfizer. M.R.W. has received educational grants from Pfizer for experimental studies in pulmonary and systemic hypertension in Kyrgyzstan. Received in original form November 25, 2004; accepted in final form March 18, 2005
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
